samstevens

speed :)

broadly, Liquid focuses on picking architectures that have really high throughput/low latency that are close to full attention (slow) in terms of expressivity, and then we do the data work to close any remaining gap. so the big goal is to make it very very fast to encode data without any loss in quality.

my intuition is that cross modality improvements do not really show up in practice, and especially not in small models

unfortunately it's not a strong multi-speaker/diarized output audio model. It's a future goal but it doesn't have a timeline. If you have the hardware, I think fine-tuning might be possible.

interesting, I didn't see it on the playground or on my local llama-server.

SFT -> RL. In our fine-tuning notebooks, we typically say that rollouts that go past the max context length (too much reasoning) get 0 reward, so there is slight pressure for shorter reasoning, but I don't think we have strong regularization. I think the agreed-upon recipe is to first make the model smarter (longer reasoning), then teach it to reason faster/denser.

the playground is bf16 or fp8 depending on the cloud GPU for the playground. I didn't have any issues with cows saying moo but I suspect it's just LLM non-determinism :)

definitely would not recommend using it for SWE-style tasks. But chaining a couple bash calls together to find a file? no problem

Can you describe more in terms of what moderation you want? I can say with confidence that we don't train for the typical adult-only content you can see on some "de-safe'd" models, but it's not clear to me exactly what you're asking about. We can also discuss over email if you prefer.

unfortunately the models are still stronger without extra modalities. So LFM3-1.7B is better at pure text tasks than all other models. the LFM3-VL-2B is better at text/vision tasks than the Omni, the Audio model is a stronger ASR/TTS model than the Omni, etc.

It's a goal of ours to continue unifying training across modalities so that there is truly one model, instead of just this shared pre-trained backbone.

It all depends on your hardware and throughput requirements. Suppose

1. I were building for my work laptop (M4 Max, 36GB RAM)

2. I had a batched use cases (transcribing + making action items for all meetings in a day)

I would use the LFM3-Audio-2B for ASR (audio -> text transcript) and then LFM3-1.7B for extracting particular action items and adding them to my tool of choice (Linear, Jira, Notes app, etc).

If you can fine-tune (100+ labeled samples, and a GPU), you likely can fine-tune LFM3-Audio-2B for audio -> action items in a single step.

You might also consider the decision models, d1-2B is very strong and good for calibrated classification. Would be happy to share links/resources if you're interested.

quantization is always a problem; what quant were you using?

gliner-style models are absolutely crazy, I learned about them a couple weeks ago when a co-worker fine-tuned an LFM encoder for promptable classification. as far as I know, you don't need to fine-tune; you can just prompt for classification. My co-worker's demo is here: https://huggingface.co/spaces/LiquidAI/prompt-routing

what does your home assistant do? a general siri/alexa-style model? or something else?

I think mlx-vlm should support both the Audio and the Omni model: https://github.com/Blaizzy/mlx-vlm/ and there should be native MLX exports for both here:

- https://huggingface.co/LiquidAI/LFM3-Audio-2B-MLX-8bit

- https://huggingface.co/LiquidAI/LFM3-Omni-2B-MLX-8bit

Yes! LFM3-VL-2B is a very strong OCR model for its size, especially given how general a model it is. You can also fine-tune it with a specific output format and it will dramatically improve. We have a Google Colab here: https://colab.research.google.com/github/Liquid4All/cookbook...

Which specific model? and are tools available? I would say that there is lots of RL with tool use, but that the reasoning traces are not always perfect. In general, models think about things that might be wrong, but will recover from this poor line of thinking.

Depends on what you care about.

If you want to do local classification/image processing, I think the d1-2B is amazing, it can run on 1GB of unified RAM.

Now that it's released and I have weekends off again, I'm going to play with the VL-2B model and the d1-2B and see if I can prompt it to control a small Lego robot arm like this https://bricksrl.github.io/ProjectPage/

If you like agents, I think the 1.7B is very strong for its size and can be used to automate small stuff like classifying emails, cleaning your downloads folder, etc.

It can also be fine-tuned for specific use cases with SFT/DPO/GRPO. The blog post talks about fine-tuning for wind-turbine classification. If there's things you want to automate, but GPT-6-Luna is not strong enough, and bigger models are too expensive, you might fine-tune one of the LFM3 models and use it locally for repetitive tasks.

I worked on this release, happy to answer questions/chat about local AI vs cloud frontier models.

in my experience it's stronger for terminal use, like chaining bash calls, monitoring background tasks, etc. rather than for straight up coding/swe work. Good with web search and file search rather than just editing software.

I worked on this model at Liquid and will be checking in to answer any questions/discussions, both about this vision model and LiquidAI more broadly.

hey I’m just spinning up in ssl birdsong models (BirdMAE, SongMAE, etc) can you share any resources? My email is [email protected], would love to read your work.

I’m really excited to see some more open SAE work! The engineering effort is non trivial and I’m going to check out your dataloading code tomorrow. You might be interested in an currently in-progress project of mine to train SAEs on vision models: https://github.com/samuelstevens/saev

Very odd that they don't do linear probing on the model features and measure ImageNet validation accuracy. In general, this paper is missing comparisons to existing vision models on existing benchmarks.

The visual prompting is a neat trick and it's great to see scale continue to work, but without comparisons to other models or releasing the code/weights, I don't think this strategy is going to be competitive.

This looks incredible! I've bookmarked this to check out later, but as an immediate question: do any of these support Apple Music/Spotify as a backend for downloading music? I'd prefer to stick with a monthly subscription for access to all music if at all possible.

Edit: also, you have an incredibly pretty website.

We just have different music listening processes. My process used to be supported (browsing my Library by Artist, Album, Genre, etc.) and got pushed to the side to support a process more like yours (Radio, Apple's curated mood playlists, etc.). I wish Apple's designers didn't limit my process in order to make room for other peoples'.